Recovering From Penguin: The Manual Link Cleanup

Posted by Sean on May 16, 2012

A few weeks on from Penguin and the question every affected client is asking is the same: how do we get back? I’ve been living inside this problem since late April and I want to write down what recovery actually looks like right now, because there’s a lot of comforting nonsense being sold about it.

The uncomfortable part first. There is no tool that makes bad links disappear. If manipulative links are dragging your site down, the only way to deal with them is to get them taken down, one at a time, by the people who host them. That means finding them, working out who controls each one, and asking, politely, for their removal.

The process, such as it is, goes like this:

  • Pull your full link profile. I run Open Site Explorer and Majestic together, because neither one sees everything; between them you get closer. Export the lot into a spreadsheet.
  • Go through it by hand and mark the links you’d be embarrassed to show a Google engineer. Paid directory listings, spun-article links, footer links on sites in unrelated languages.
  • For each bad link, find a contact. This is where the days vanish. Half these sites have no contact page. You end up doing WHOIS lookups, guessing at admin@ addresses, filling in grim little contact forms.
  • Email the webmaster and ask them to remove the link. Keep a record of every request; the date, the address, exactly what you said.

That last step sounds simple and it is soul-destroying. A good number of the sites won’t reply. Some will reply asking for money to take down a link you never wanted. Some no longer have a working owner at all. You’ll spend a fortnight chasing removals and come out the far side having cleared perhaps half of what you flagged, the rest sitting there because nobody’s home.

Once you’ve removed as much as you can, and only if you had a manual penalty notice in Webmaster Tools, you file a reconsideration request. This bit matters. Reconsideration only helps if a human at Google took manual action against you. A lot of Penguin hits appear to be algorithmic; no message, no manual flag. For those, a reconsideration request does nothing at all, and you’re simply waiting for the algorithm to run again and re-assess you. Telling the two apart is the first thing to check, and plenty of people are burning effort on reconsideration requests for an algorithmic drop that will only lift when Google refreshes the data.

When you do file one, show your working. Google’s reviewers want evidence that you understand what went wrong and have genuinely tried to fix it. That spreadsheet of removal emails, dated, is the thing to point at. “We identified X problem links, removed Y, and here are the ones we couldn’t reach and what we tried.” Vague apologies get form-letter rejections.

I’m three weeks into this on a couple of sites and the plain state of it is tedious, slow, and only partly within your control. Keep the spreadsheet tidy. You’ll be living in it for a while.